Bug 215286 – DFG should always run CFG Simplification after Constant Folding.

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Keith Miller]

DFG should always run CFG Simplification after Constant Folding.

[Keith Miller]

Created attachment 406199 [details]
Patch

[Robin Morisset]

Comment on attachment 406199 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=406199&action=review

> Source/JavaScriptCore/dfg/DFGCFGSimplificationPhase.cpp:-52
> -        // FIXME: We should make this work in SSA. https://bugs.webkit.org/show_bug.cgi?id=148260

If this is no longer a FIXME, please close the corresponding bug on bugzilla.

> Source/JavaScriptCore/dfg/DFGValidate.cpp:-814
> -            bool didSeeExitOK = false;

Why did you remove this validation? Can you explain it in the Changelog if it is on purpose?

[Keith Miller]

*** Bug 148260 has been marked as a duplicate of this bug. ***

[Keith Miller]

Comment on attachment 406199 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=406199&action=review

>> Source/JavaScriptCore/dfg/DFGCFGSimplificationPhase.cpp:-52
>> -        // FIXME: We should make this work in SSA. https://bugs.webkit.org/show_bug.cgi?id=148260
> 
> If this is no longer a FIXME, please close the corresponding bug on bugzilla.

Done.

>> Source/JavaScriptCore/dfg/DFGValidate.cpp:-814
>> -            bool didSeeExitOK = false;
> 
> Why did you remove this validation? Can you explain it in the Changelog if it is on purpose?

Because it's unnecessary and fails when you merge two blocks where the latter block has a Phi. I'll add a comment to the ChangeLog.

[Keith Miller]

Created attachment 406310 [details]
Patch

[Radar WebKit Bug Importer]

<rdar://problem/67087402>

[EWS]

Committed r266101: <https://trac.webkit.org/changeset/266101>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 406310 [details].